feature-flags

Design percentage rollouts and A/B tests with sample size calculations.

18|1|Updated Apr 17, 2026
One-click install
npx skills add https://github.com/kid-sid/codex-spellbook --skill feature-flags-kid-sid
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: feature-flags
Source: https://github.com/kid-sid/codex-spellbook/tree/main/skills/feature-flags
Command: npx skills add https://github.com/kid-sid/codex-spellbook --skill feature-flags-kid-sid

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill eliminates the risk of unvalidated feature releases, inconsistent feature flag governance, and statistically invalid A/B experiments that waste engineering resources and harm user experience.

Core Features & Use Cases

  • End-to-End Flag Implementation: Native support for LaunchDarkly, Unleash, and OpenFeature with SDK setup, evaluation context configuration, and safe default handling for offline or uninitialized states.
  • Rollout & Experiment Design: Percentage-based and ring-based rollouts, sticky bucketing, mutual exclusion for concurrent experiments, and pre-registered sample size calculation for statistically valid A/B tests.
  • Lifecycle Governance: Standardized flag naming conventions, stale flag cleanup detection, three-phase flag migration workflows, and guardrail metric definition to prevent performance or revenue regressions. Use case example: A payments team can use this skill to roll out a new checkout flow to 5% of users, run a controlled A/B test against a control group, and safely migrate to the new code once the experiment meets success criteria.

Quick Start

Use the feature-flags skill to design a percentage rollout and A/B test for the new user onboarding flow, including sample size calculation and guardrail metric setup.

Frequently Asked Questions about feature-flags

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I implement a safe percentage rollout for a new feature using LaunchDarkly or Unleash?

To implement a safe percentage rollout, configure SDK initialization patterns and evaluation context for LaunchDarkly or Unleash, ensuring safe default handling for offline states. This enables gradual ring-based delivery to minimize release risk.

What is sticky bucketing and how does it work for A/B testing?

Sticky bucketing ensures users remain in the same experiment variation across sessions by maintaining consistent evaluation context. This prevents metric contamination and ensures statistically valid A/B test results.

Can I use OpenFeature to standardize feature flag evaluation across multiple services?

Yes, you can use OpenFeature to standardize feature flag evaluation across services. It provides vendor-agnostic SDK initialization and evaluation context configuration, ensuring consistent flag lifecycle governance and safe default handling.

How do I calculate minimum sample size for a statistically valid A/B experiment?

Calculate minimum sample size for a statistically valid A/B experiment by pre-registering sample size requirements and defining guardrail metrics. This prevents premature conclusions and protects against performance or revenue regressions.

What is the best way to manage stale feature flags and govern their lifecycle?

The best way to manage stale feature flags is applying standardized naming conventions and stale flag cleanup detection within a three-phase flag migration workflow. This ensures maintainable feature delivery across services.

How do I prevent concurrent A/B tests from interfering with each other?

Prevent concurrent A/B tests from interfering by configuring mutual exclusion rules and sticky bucketing logic. This isolates experiment variants and maintains the statistical validity of multivariate tests.